This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: RFC: allow breakpoints on labels
- From: Tom Tromey <tromey at redhat dot com>
- To: Michael Snyder <msnyder at vmware dot com>
- Cc: "gdb-patches\ at sourceware dot org" <gdb-patches at sourceware dot org>
- Date: Fri, 21 Aug 2009 16:14:59 -0600
- Subject: Re: RFC: allow breakpoints on labels
- References: <m38whe5rul.fsf@fleche.redhat.com> <4A8D8A26.7010303@vmware.com>
- Reply-to: Tom Tromey <tromey at redhat dot com>
>>>>> "Michael" == Michael Snyder <msnyder@vmware.com> writes:
Michael> 1) Labels will be in a separate name-space from other symbols
Michael> (functions)?
Yes. Here I just reused some existing gdb infrastructure, which appears
to mostly be unused.
You can't even "print label" at the moment, due to this.
It might be nice to implement the GCC address-of-label extension, though
this doesn't seem to be extremely useful.
Michael> 2) The search order will find functions first, and only if
Michael> there is no conflicting function, find a label? What if
Michael> there's a function and a label of the same name, and user
Michael> wants to specify the label?
Yes, functions are searched first. There's no way to pick the label in
this case.
Michael> 3) There's no danger of our confusing a label with a
Michael> function somewhere downstream, and doing something silly
Michael> such as applying prologue_skip or skip_trampoline to a label?
I don't think so, but I'm really not sure how I would be sure.
Tom